Building the Future of AI
& Data Infrastructure
We architect, build, and deploy production-grade AI systems and data platforms — from real-time streaming pipelines processing terabytes daily to advanced Agentic AI and RAG architectures that transform how enterprises leverage intelligence.
End-to-End AI & Data
Platform Engineering
We don't just consult — we architect from scratch, build, and deploy. From designing distributed data pipelines that handle >1 TB/day streaming with Kafka, Spark, and Flink, to engineering Agentic AI systems with advanced RAG and LLM orchestration — we deliver infrastructure that scales.
&
WHAT WE HAVE ALREADY ACHIEVED :
Comprehensive AI & Data Services
Every service is delivered with production-grade rigor — from architecture to deployment.
AI Consultancy
Strategic guidance on AI adoption, architecture decisions, and roadmap planning tailored to your business goals.
AI Architecture from Scratch
Complete AI infrastructure design — from data ingestion to model serving, built on AWS/Databricks with scalability at core.
Agentic AI Systems
Autonomous LLM agents with tool use, error recovery loops, and multi-agent orchestration for complex workflows.
Model Training & Fine-tuning
Custom model training pipelines with HuggingFace, NVIDIA NIM, and on-device edge deployment (Gemma/LiteRT).
RAG Architectures
Advanced Retrieval-Augmented Generation with two-stage retrieval, cross-encoder reranking, and streaming real-time ingestion.
Data Platform Setup
End-to-end data platforms — pipelines to analytics — using Databricks, AWS, Spark, Flink, Kafka, Hive/Hudi, and Presto.
Built for Production,
Engineered for Scale
We bring a rare blend of deep data engineering and cutting-edge AI expertise — ensuring your systems aren't just prototypes, but production-hardened infrastructure.
High-Throughput Streaming
Pipelines processing 1 TB/day with Kafka, Spark, and Flink — built for real-time enterprise workloads.
Data Reliability & Quality
85%+ test coverage, robust ETL/ELT frameworks, and data mapping that standardizes heterogeneous flows.
Cloud-Native & Cost-Optimized
AWS/Azure migration expertise that reduced infrastructure costs by 25% while improving latency by 30%.
IEEE-Published Research
Published NLP/ML research with novel data augmentation pipelines — bridging academia and industry.
PUNEET SARAN aKa S19
Data Engineer | AI/ML Engineer
🧠 Crafting intelligent systems that don't just process data—they understand it.
From streaming terabytes in real time to architecting autonomous AI agents, I bridge the gap between raw infrastructure and transformative GenAI. My code scales to petabyte pipelines, my models speak in vectors, and my architectures are battle-tested across startups, EV giants, and enterprise AI labs. IEEE-published, cloud-optimized, and relentlessly focused on data reliability—I engineer solutions where every byte counts.
📍 Bangalore, India | 📧 saran19.work@gmail.com | 📱 +91-9256186314
Work Experience
Kindrix Group — Data Engineer
June 2024 – Dec 2025
Architected scalable AI data platforms with production-grade RAG pipelines (LangChain, Vector DBs, LLMs) for enterprise document parsing and B2B client workloads.
Sapper AI (Zyient) — DPE (Backend)
Jul 2023 – Nov 2023
Engineered automated ETL/ELT document pipelines to feed LLM-powered RAG systems, reducing manual review effort and improving answer accuracy for 3 enterprise clients. Cut unstructured data ingestion time by 40% via robust data mapping framework.
Ola Electric — SDE I
Jun 2022 – Jul 2023
Engineered 1 TB/day streaming pipelines for EV sensor data; optimized latency by 30% and reduced infra costs by 25% during cloud migration to AWS/Azure.
Samsung — PRISM Developer
Jul 2020 – Apr 2021
Developed ML/NLP models for Bixby voice trigger recognition; published novel NER data augmentation pipeline in IEEE Xplore (2023).
Monktree Education Pvt Ltd — App Developer Intern
May 2020 – Jul 2020
Enhanced UI and functionality of a financial mobile app (tax calculator); improved user experience and reliability through testing and bug fixes.
Recent Projects
🤖 Agentic AI — Wannabae
Low-latency NLP engine computing 5-dimensional personality vectors in under 2 seconds using NRC-VAD lexicon and NVIDIA LLM fallback.
💼 Job_Ops — Autonomous LLM Command Center
Modular agentic system automating web scraping, structured extraction, and ATS-optimized PDF generation with Groq's Llama-3.
🎨 Movator — Multimodal Generative Pipeline
Dual-agent AI orchestrating Llama 3.2 and Flux.1 with autonomous error-recovery loops for continuous pipeline execution.
📡 Kafka-RAG — Real-Time Document Intelligence
Live streaming RAG pipeline with two-stage retrieval (ChromaDB + NVIDIA Nemotron reranking) feeding Groq Llama-3; evaluated via RAGAs.
Education
B.Tech in ECE
VIT University, 2018–2022 | CGPA: 8.78
MA Geography
2024–2026
Let's Build Together
Ready to architect your AI infrastructure or data platform? Reach out — let's discuss how we can transform your data into intelligence.